ops: Disk tier occupancy on /metrics
CI / vulncheck (pull_request) Successful in 19s
CI / check-and-test (pull_request) Successful in 47s
CI / vulncheck (push) Successful in 16s
CI / check-and-test (push) Successful in 46s
Release Tag / release (push) Successful in 14s

Operators could see attach-ready and capacity-pressure events but not how
full the configured disk (or memory) tier was without reading filesystems.
Expose size/capacity gauges and disk_cache_full_ratio next to disk_tier_ready.
Capacity is a config read, so it stays available while attach is pending.
This commit was merged in pull request #53.
This commit is contained in:
2026-09-09 16:06:12 +00:00
parent a3ea4806a7
commit 50ca0a071c
5 changed files with 240 additions and 2 deletions
+61
View File
@@ -100,6 +100,9 @@ func TestWriteTextPrometheusExposition(t *testing.T) {
TotalBytesSaved: 50,
MemoryCacheSize: 8,
DiskCacheSize: 16,
MemoryCacheCapacity: 8,
DiskCacheCapacity: 32,
DiskCacheFullRatio: 0.5,
DiskTierReady: 1,
Uptime: 3 * time.Second,
}
@@ -120,6 +123,7 @@ func TestWriteTextPrometheusExposition(t *testing.T) {
}
gauges := []string{
"hit_rate", "avg_response_time_ms", "memory_cache_size", "disk_cache_size",
"memory_cache_capacity", "disk_cache_capacity", "disk_cache_full_ratio",
"disk_tier_ready", "uptime_seconds",
}
for _, name := range counters {
@@ -138,6 +142,15 @@ func TestWriteTextPrometheusExposition(t *testing.T) {
if !strings.Contains(body, "disk_tier_ready 1\n") {
t.Errorf("missing disk_tier_ready 1 sample: %q", body)
}
if !strings.Contains(body, "memory_cache_capacity 8\n") {
t.Errorf("missing memory_cache_capacity 8 sample: %q", body)
}
if !strings.Contains(body, "disk_cache_capacity 32\n") {
t.Errorf("missing disk_cache_capacity 32 sample: %q", body)
}
if !strings.Contains(body, "disk_cache_full_ratio 0.5000\n") {
t.Errorf("missing disk_cache_full_ratio 0.5000 sample: %q", body)
}
if !strings.Contains(body, "range_cache 3\n") {
t.Errorf("missing range_cache 3 sample: %q", body)
}
@@ -146,6 +159,54 @@ func TestWriteTextPrometheusExposition(t *testing.T) {
}
}
func TestDiskCacheFullRatioInGetStats(t *testing.T) {
t.Parallel()
cases := []struct {
name string
size int64
capacity int64
wantRatio float64
wantCapacity int64
}{
{"no disk (capacity 0)", 0, 0, 0, 0},
{"zero size with capacity", 0, 1024, 0, 1024},
{"half full", 512, 1024, 0.5, 1024},
{"exact full", 1024, 1024, 1, 1024},
{"size above capacity clamps to 1", 2048, 1024, 1, 1024},
}
for _, tc := range cases {
tc := tc
t.Run(tc.name, func(t *testing.T) {
t.Parallel()
m := NewMetrics()
m.SetDiskCacheSize(tc.size)
m.SetDiskCacheCapacity(tc.capacity)
st := m.GetStats()
if st.DiskCacheCapacity != tc.wantCapacity {
t.Fatalf("DiskCacheCapacity=%d, want %d", st.DiskCacheCapacity, tc.wantCapacity)
}
if st.DiskCacheFullRatio != tc.wantRatio {
t.Fatalf("DiskCacheFullRatio=%v, want %v", st.DiskCacheFullRatio, tc.wantRatio)
}
if st.DiskCacheFullRatio < 0 || st.DiskCacheFullRatio > 1 {
t.Fatalf("DiskCacheFullRatio=%v outside [0,1]", st.DiskCacheFullRatio)
}
})
}
// Memory capacity is a plain passthrough, and capacity survives Reset
// (re-derived by GetMetrics, like MemoryCacheSize/DiskTierReady).
m := NewMetrics()
m.SetMemoryCacheCapacity(4096)
if got := m.GetStats().MemoryCacheCapacity; got != 4096 {
t.Fatalf("MemoryCacheCapacity=%d, want 4096", got)
}
m.Reset()
if got := m.GetStats().MemoryCacheCapacity; got != 4096 {
t.Fatalf("MemoryCacheCapacity=%d after Reset, want 4096 (config snapshot, like size gauges)", got)
}
}
func assertHelpType(t *testing.T, body, name, typ string) {
t.Helper()
help := "# HELP " + name + " "